Use Virtual Machine Prioritization settings to specify the order in which VMs migrate, shut down and power on.
|
VM Prioritization is only available for hosts managed by vCenter Server. NOTE: If vCenter Server is unavailable at the time of the shutdown, PowerChute will still use the order determined by the priority group to shut down VMs when it connects to each ESXi host. vCenter Server must be available for Dell VxRail cluster shut down. |
VM Prioritization is configured in the main PowerChute interface and is
disabled by default.
Enable VM Prioritization screen
To enable VM Prioritization, select the Enable VM Prioritization checkbox. Four options display:
Virtual Machines can be grouped into five priority groups – High, Medium, Low, Group 1 and Group 2. When VM Prioritization is enabled, an inventory view of the datacenter, clusters, vCenter Server Appliance (VCSA), VMs and vApps appears on the left. On the right, the High, Medium, Low, Group 1 and Group 2 priority groups are listed.
Datacenter inventory view and VM Priority Groups
You can assign a VM/vApp to a priority group by clicking on a VM/vApp on the left hand side and dragging it to a priority group on the right.
Move VMs/vApps between priority groups by dragging them from one group
to another. To remove a VM/vApp from a priority group, select the VM/vApp
and click the Remove button. Any VM/vApp in
the inventory that is not assigned to a priority group is considered to
be Un-prioritized.
|
The vCenter Server VM can be added to a priority group if there are special VMs that need to be shut down after vCenter Server and started before it. For example, Platform Services Controller VM, Active Directory Controllers. |
|
The inventory view on the left is populated with VMs that are part of the same HA cluster as any Host protected by a UPS specified on the Host Protection page. VMs on Standalone Hosts managed by vCenter are also present.
|
 |
For PowerChute to correctly identify clusters, VMs and vApps in a datacenter, their names must be unique. It is recommended that a separate instance of PowerChute is used for each datacenter. If one instance of PowerChute is used to protect hosts in multiple datacenters, it will not be able to identify VMs, vApps or clusters that have the same name, in different datacenters. In each datacenter in which PowerChute is used to protect hosts, make sure that VMs, vApps and Clusters each have a unique name. In a Dell VxRail configuration, it is recommended that a separate PowerChute instance is deployed for each VxRail cluster. |
You can configure durations for each priority group, for VM Migration, VM Shutdown and VM Startup.
For Priority Groups:
VM Migration Duration sets the time allowed for all VMs in the Priority Group to migrate to another healthy host in the cluster. This also represents the delay between migrating VMs for each Priority Group.
|
The VM Shutdown Duration can be set to 0 seconds for any priority group to handle any special VMs that need to be shut down later in the shutdown sequence via a command file or SSH action. This is relevant for VMs running on the Management Host in a Dell VxRail stretched cluster configuration. If you set a 0 second shutdown duration for a priority group, the shutdown action will be skipped for all VMs in that priority group. |
Set VM Operation durations - VM Migration, Shutdown and Startup
You can set durations for the VMs/vApps in the High, Medium, Low, Group 1 and Group 2 priority Groups, and for VMs/vApps that are Un-prioritized. When VM Prioritization is enabled for the first time, High, Medium, Low, Group 1 and Group 2 durations will have a default value of 0.
Durations for un-prioritized VMs/vApps are automatically set to the global duration values for VM Migration, VM Shutdown and VM Startup, as configured on the Virtualization Settings page.
|
If vApps are added to a priority group, they will not be shut down if vCenter Server is offline and no duration is specified for Un-prioritized VMs/vApps. To ensure your vApps gracefully shut down, provide a shutdown duration for the Un-prioritized priority group. |
VM Migration and VM Startup durations can be configured to have a value of zero, to skip the migration or startup of VMs/vApp when a critical event is resolved. If VM Migration/VM Startup is set to zero for a priority group, VMs/vApps that do not migrate will be shut down, and will not start up following the shutdown, when the critical event is resolved.
A warning will display in the VM Prioritization screen if a VM Shutdown duration is set to zero for any priority group.
When VM Prioritization is enabled, the VM Migration, VM Startup and VM Shutdown durations set on the Virtualization Settings page automatically match the sum of the respective High/Medium/Low//Group 1/Group 2/Un-prioritized durations set on the VM Prioritization page.
For example, if VM Shutdown durations set on the VM Prioritization screen are as follows:
then the VM Shutdown duration on the Virtualization Settings page is automatically set to:
90 + 60 + 60 + 30 + 30 + 30 = 300 seconds
Virtualization Setting page - VM Shutdown Duration
If Advanced UPS configuration is used, the durations for each UPS Setup are also changed to reflect the sum of the respective priority group durations for VM Migration, Startup and Shutdown.
On the Virtualization Settings page when VM Prioritization is enabled, VM Migration, Startup and Shutdown durations become read-only and are automatically populated, as described above. If VM Prioritization is disabled, the durations retain the values set by VM Prioritization, but become editable once more.
Operations such as VM Migration, VM Shutdown and VM Startup can be performed for VMs in a priority group.
For VM Migration and VM Startup, the order in which priority groups are processed is as follows:
For VM Shutdown the order in which priority groups are processed is:
The sequences below describe the VM Migration, Shutdown and Startup Sequence for all UPS Configurations.
In this scenario:
VM Migration is enabled
DRS is disabled
Wait for all Hosts Online is enabled
VM Shutdown is enabled
VM Startup is enabled
A UPS Critical event occurs and PowerChute starts a maintenance mode task on each protected host. First the High priority group VMs migrate, in parallel. When all High priority Group VMs have migrated, the Medium priority VMs migrate in parallel, followed by the Low, Group 1 and Group 2 priority groups, and finally the Un-prioritized VMs migrate.
|
PowerChute will proceed to migrate the next group of VMs/vApps when the duration for the priority group has elapsed. |
When the VM migration duration elapses, any VMs that have not been migrated will be shut down in the VM Shutdown sequence.
PowerChute begins sequenced VM/vApp Shutdown. First, the un-prioritized VMs/vApp are shut down sequentially. After the duration for un-prioritized VMs/vApps elapses, the Group 2 priority VMs/vApps are shut down, followed by the Group 1 priority VMs/vApps, the Low priority VMs/vApps, the Medium priority VMs/vApps and finally the High priority VMs/vApps are shut down. When all priority group VMs/vApps have been shut down, PowerChute shuts down the vCenter Server VM if vCenter Server is running on a VM and the vCenter Server VM is not part of a priority group. PowerChute then shuts down the hosts and the physical PowerChute server. In a Dell VxRail configuration, PowerChute sends a request to the NMC to shut down the VxRail cluster after a configurable delay. PowerChute then shuts itself down.
When the UPS critical event is resolved and the hosts and physical machine running PowerChute are powered back on, the hosts are taken out of maintenance mode and the vCenter Server VM is started if the vCenter Server VM has not been added to a priority group. In a Dell VxRail configuration, VxRail handles taking the hosts out of maintenance mode and re-starting vCenter Server VM and VxRail Manager VMs. PowerChute begins to start the High priority VMs/vApps with respect to their startup duration. When the High priority startup duration elapses, the Medium priority VMs/Vapps are started, followed by the Low priority VMs/vApps, the Group 1 priority VMs/vApps, the Group 2 priority VMs/vApps, and finally, the un-prioritized VMs/vApps are started.
In Advanced UPS configuration, ESXi hosts and equipment may be protected by separate UPS devices.
In this example, UPS 2 may experience a critical event at a different time to UPS 3. If the UPS Setups for Host A and B have VM Prioritization enabled, the VM operations that are performed on the priority groups of the two hosts will occur at different times. For example, if UPS 2 experiences a critical event 2 minutes before UPS 3, the High priority VMs on Host B may be shut down at the same time as the Low priority VMs on Host A. VM operations on priority groups occur at the host level.
This also applies in a multi-site scenario where hosts in two geographical locations, Site A and Site B, are powered by two different UPS devices. As the critical event may not occur at the same time at each location, if Site A experiences a critical event first, it is possible to have High priority VMs shutting down in site B at the same time as Low priority VMs shut down at site A.
